A noise-tolerant approach to strain estimation in phase-sensitive optical coherence elastography, robust to decorrelation distortions, is discussed. The method is based on evaluation of interframe phase-variation gradient, but its main feature is that the phase is singled out at the very last step of the gradient estimation. All intermediate steps operate with complex-valued optical coherence tomography (OCT) signals represented as vectors in the complex plane (hence, we call this approach the 'vector' method). In comparison with such a popular method as least-square fitting of the phase-difference slope over a selected region (even in the improved variant with amplitude weighting for suppressing small-amplitude noisy pixels), the vector approach demonstrates superior tolerance to both additive noise in the receiving system and speckle-decorrelation caused by tissue straining. Another advantage of the vector approach is that it obviates the usual necessity of error-prone phase unwrapping. Here, special attention is paid to modifications of the vector method that make it especially suitable for processing deformations with significant lateral inhomogeneity, which often occur in real situations. The method's advantages are demonstrated using both simulated and real OCT scans obtained during reshaping of a collagenous tissue sample irradiated by an IR laser beam producing complex spatially inhomogeneous deformations.

In compressional optical coherence elastography, phase-variation gradients are used for estimating quasistatic strains created in tissue. Using reference and deformed optical coherence tomography (OCT) scans, one typically compares phases from pixels with the same coordinates in both scans. Usually, this limits the allowable strains to fairly small values <10?4 to 10?3, with the caveat that such weak phase gradients may become corrupted by stronger measurement noises. Here, we extend the OCT phase-resolved elastographic methodology by (1) showing that an order of magnitude greater strains can significantly increase the accuracy of derived phase-gradient differences, while also avoiding error-phone phase-unwrapping procedures and minimizing the influence of decorrelation noise caused by suprapixel displacements, (2) discussing the appearance of artifactual stiff inclusions in resultant OCT elastograms in the vicinity of bright scatterers due to the amplitude-phase interplay in phase-variation measurements, and (3) deriving/evaluating methods of phase-gradient estimation that can outperform conventionally used least-square gradient fitting. We present analytical arguments, numerical simulations, and experimental examples to demonstrate the advantages of the proposed optimized phase-variation methodology.

No abstract

Quantitative mapping of deformation and elasticity in optical coherence tomography has attracted much attention of researchers during the last two decades. However, despite intense effort it took ~15 years to demonstrate optical coherence elastography (OCE) as a practically useful technique. Similarly to medical ultrasound, where elastography was first realized using the quasi‐static compression principle and later shear‐wave‐based systems were developed, in OCE these two approaches also developed in parallel. However, although the compression OCE (C‐OCE) was proposed historically earlier in the seminal paper by J. Schmitt in 1998, breakthroughs in quantitative mapping of genuine local strains and the Young's modulus in C‐OCE have been reported only recently and have not yet obtained sufficient attention in reviews. In this overview, we focus on underlying principles of C‐OCE; discuss various practical challenges in its realization and present examples of biomedical applications of C‐OCE. The figure demonstrates OCE‐visualization of complex transient strains in a corneal sample heated by an infrared laser beam.

In this paper, we point out some practical obstacles arising in realization of compressional optical coherence elastography (OCE) that have not attracted su±cient attention previously. Speci¯-cally, we discuss (i) complications in quanti¯cation of the Young modulus of tissues related to partial adhesion between the OCE probe and soft intervening reference layer sensor, (ii) distorting in°uence of tissue surface curvature/corrugation on the subsurface strain distribution mapping, (iii) ways of signal-to-noise ratio (SNR) enhancement in OCE strain mapping when periodic averaging is not realized, and (iv) potentially signi¯cant in°uence of tissue elastic nonlinearity on quanti¯cation of its sti®ness. Potential practical approaches to mitigate the e®ects of these complications are also described.